Parrot 2.11.0 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Onward and upward with Google Code-In | Please test rakudo with bleeding edge parrot! | merge html_cleanup and embed_api2
Set by moderator on 23 December 2010.
00:01 Khisanth joined
dalek rrot: 8138ebd | jkeenan++ | src/dynpmc/Rules.in:
Remove include/pmc/pmc_default.h and include/pmc/pmc_continuation.h throughout, as they are already included in $DYNPMC_H_FILES. t/codingstd/checkdepends.t should now PASS.
 00:01
NotFound msg plobsing winxed r722 may be what you want 00:03
aloha OK. I'll deliver the message.
dalek rrot: c0b71ed | (Matt Rajca)++ | src/embed/pmc.c:
Use Parrot_pmc_box_string to box strings

Signed-off-by: Christoph Otto <christoph_github@mksig.org>
rrot: 375c610 | (Matt Rajca)++ | / (6 files):
Unified boxing of strings

Signed-off-by: Christoph Otto <christoph_github@mksig.org>
plobsing NotFound: pong 00:06
NotFound plobsing: see winxed r722
plobsing will try. thanks. 00:07
NotFound Uhhh... guys, are you sure that that change to the box opcode makes sense?
cotto_work NotFound: good catch. 00:08
I'm a little disappointed that we don't have a test for that. 00:09
NotFound It passes all tests? If it does, we need a test for that"
!
cotto_work wins
NotFound TouchƩ
00:09 Khisanth left
plobsing shouldn't string-box use HLL mapping? 00:09
cotto_work Matt_: ping
plobsing: yes
NotFound been reading Les trois mousquetaires
cotto_work Matt_: can you make Parrot_pmc_box_string use Parrot_hll_get_ctx_HLL_type(interp, enum_class_String) for the type and add a test to make sure that hllmapped strings are used for boxing? 00:10
Matt_ cotto_work: will do
cotto_work excellent. Thanks.
Matt_ I ran `make test` and everything passed
NotFound MMmm.... but the function is for embedding, isn't it?
The embedder should not have 'current' HLL. 00:11
Matt_: we don't have enough tests, that's why we want to improve coverage. 00:12
Yuki`N cotto, is there any work for me to be doing? I'm kinda just stagnating right now. 00:13
cotto_work NotFound: do you think it's a bad idea to use the hllmapped type in an embedding function?
Yuki`N: nothing on the gci list grabs you? 00:14
Yuki`N No, not really.
There's a task for adding gcno and etc to gitignore
That's only one point though
cotto_work I need to nuke that.
kid51 took care of it.
Yuki`N And all the "test coverage" tasks are somewhat outdated.
I'm looking to work with some of the new embedding API 00:15
Hopefully if you can point me in the right direction I can help push to get the api ready to be merged back into master.
Matt_ cotto_work: should I proceed with changing Parrot_pmc_box_string to use Parrot_hll_get_ctx_HLL_type?
Yuki`N Whiteknight seems to have ignored me every time I bugged him about it though. :3
NotFound cotto_work: what will be the current HLL? The HLL belongs to a invokable thing or a context, doesn't make sense to the outside world. 00:16
cotto_work NotFound: is there any harm in it? By default it'll just use the normal string pmc.
NotFound cotto_work: confusion. 00:17
...will be my epitaph
00:17 Khisanth joined
cotto_work I don't see the harm but this is whiteknight's domain. I'd ask him and go with what he says. 00:20
plobsing I was confused I thought that this was touching the box string op. that needs hll map.
NotFound cotto_work: agrred 00:21
agreed
cotto_work plobsing: part of the confusion is that that's what it's doing.
The box op clearly needs hllmap-awareness. Whether it belongs in the embed api is an open question. 00:22
plobsing I'd argue not. 00:23
00:24 kid51 is now known as kid51_at_dinner
NotFound Apparently we have a test for that ? 00:27
ok 26 - string boxed to appropriate type for HLL from reg isa MyString 00:28
Matt_ Just modified Parrot_pmc_box_string to use Parrot_hll_get_ctx_HLL_type and all tests run fine 00:29
Any next steps for me?
00:29 theory joined 00:35 Khisanth left
cotto_work Matt_: I'd like to know why that test (box_string, end of t/op/box.t) passes. 00:41
Yuki`N: if you can think of a task, some can write it up for you. atm my brain is in too many places to think of something. 00:42
Yuki`N Ah.
I'm not too sure what needs to be done tbh 00:43
I'll try asking whiteknight again
cotto_work Yuki`N: how do you feel about packfiles? 00:44
trac.parrot.org/parrot/wiki/PackfileTasklist
Yuki`N Umm, the bytecode container?
cotto_work that's one of the things it does
cotto_work heads home 00:46
00:56 kid51_at_dinner is now known as kid51 01:15 nwellnhof left
kid51 Here's the result of an up-to-date run of 'make quickcover' 01:26
thenceforward.net/parrot/coverage/q...erage.html
dalek rrot: 2a95633 | jkeenan++ | config/gen/makefiles/root.in:
Add src/embed/ to COVER_DIRS and QUICKCOVER_DIRS to guarantee that files in those directories created during coverage analysis are removed by cover-clean.
 01:27
kid51 So on my Linode that run of quickcover took 5:08 (including reconfiguration, make, make test and cover) 01:29
01:38 Khisanth joined 01:42 Khisanth left 02:24 kennym left
kid51 cotto: I just created 3 GCI tasks (as yet unapproved) 02:35
02:46 Yuki`N left 02:48 kid51 left
dalek m-eta-wink-kzd: 089a4bc | plobsing++ | src/ometa-compiler.dual:
use volatile to ensure values are passed up from closures

this happens eg: in destructuring binds		 02:59
m-eta-wink-kzd: cdd5f90 | plobsing++ | t/tutorial.Ī©Ī·:
eliminate another javascript-ism
 03:00
m-eta-wink-kzd: 838cb0f | plobsing++ | src/winxed-compiler.dual:
avoid aliasing between keyword 'string' tokens and tokens of type string
03:19 dngor_ is now known as dngor
dalek m-eta-wink-kzd: 95bff41 | plobsing++ | t/tutorial.Ī©Ī·:
disambiguate string concatenations

concatenations is not addition on strings! stop writting languages with this braindead mindset!		 03:43
m-eta-wink-kzd: bd25c51 | plobsing++ | t/tutorial.Ī©Ī·:
use braces to immitate correct round-trip expression compiler output
m-eta-wink-kzd: 332f3d4 | plobsing++ | src/ometa-base.winxed:
use method call in stead of attribute access
03:44 Matt_ left 04:05 Matt_ joined 04:11 Andy joined 04:20 zaur left 04:21 Khisanth joined 04:24 theory left, plobsing_ joined 04:28 plobsing left 04:29 plobsing joined 04:30 Khisanth left, plobsing_ left 04:33 theory joined 04:38 allison left 04:39 theory left 04:41 Matt_ left
cotto plobsing++ for fixing the dependencies/includes earlier today 04:42
04:42 plobsing_ joined 04:46 plobsing left 05:09 plobsing_ left 05:34 rurban_ joined 05:36 plobsing joined, rurban left 05:37 rurban_ is now known as rurban
dalek m-eta-wink-kzd: 3f810c7 | plobsing++ | src/ometa-base.winxed:
delegate to target in OMInputStreamProxy
 05:42
m-eta-wink-kzd: ea320da | plobsing++ | src/ (4 files):
hide js vs winxed join() difference behind macro
m-eta-wink-kzd: 28d4a09 | plobsing++ | src/ometa-base.winxed:
implement {get,set}_pmc_keyed()
06:01 Andy left 06:09 Khisanth joined, khisanth_ joined 06:15 Khisanth left 06:21 khisanth_ is now known as Khisanth 06:25 plobsing left 07:05 dd070 joined 07:07 dd070 left 07:51 simon_ joined
simon_ can PIR subroutines contain subroutines? 07:51
sorear No. 07:57
However, they can be associated with a parent Sub, and can then access the lexicals of the parent 07:58
simon_ oh. I'll get to that. 08:01
08:17 fperrad joined
cotto msg kid51 The gci task to change the names of coverage-related makfile targets seems really trivial. Is there something I'm missing? 08:37
aloha OK. I'll deliver the message.
cotto msg whiteknight Let me know if you want a gci org admin bit. 08:40
aloha OK. I'll deliver the message.
cotto msg aloha aloha
aloha OK. I'll deliver the message.
cotto bacek, ping 09:03
bacek cotto, pong. I'm barely here. It's almost x-mas here :) 09:05
nopaste "cotto" at 192.168.1.3 pasted "is this the codingstd change you're suggesting?" (24 lines) at nopaste.snit.ch/27433 09:06
cotto bacek, ^
bacek cotto, nope. Indent C<goto> as well 09:07
and C<return>
cotto so indent everything by +2?
bacek +4
nopaste "cotto" at 192.168.1.3 pasted "like this?" (24 lines) at nopaste.snit.ch/27434 09:08
"bacek" at 192.168.1.3 pasted "Indent" (10 lines) at nopaste.snit.ch/27435
bacek Like next one :)
cotto got it 09:09
Is that just to make the labels stick out more?
bacek yes
cotto wfm
bacek for consistency sake
cotto ok. I'm glad I asked. 09:11
bacek ok 09:13
afk # MORE BEER!
09:13 contingencyplan joined 09:25 particle left 09:26 particle joined 09:28 rfw left 09:46 bacek left 10:14 bacek joined
dalek rrot: 11dc7f6 | NotFound++ | t/pmc/exporter.t:
clean up a Exporter test and make sure it triggers the intended condition
 10:49
10:55 contingencyplan left
dalek rrot: 0b29d99 | NotFound++ | src/pmc/exporter.pmc:
change a condition that can't be triggered to an assertion
 11:00
11:10 kennym joined 11:17 dd070 joined 11:20 dd070 left 11:27 Kulag left 11:28 Kulag joined 11:34 Drossel joined 11:37 Kulag left 13:34 rurban_ joined 13:36 rurban left 13:37 rurban_ is now known as rurban 13:43 plobsing joined
plobsing NotFound++ # volatile worked 13:49
14:05 ambs joined
NotFound plobsing: good 14:15
14:17 bluescreen joined 15:12 bluescreen left, bluescreen joined
bluescreen merry xmas! 15:12
tadzik merry! 15:18
15:25 fbrito joined 15:31 fbrito left 15:40 Matt_ joined 15:46 fperrad left 15:49 cottoo joined, fbrito joined 15:54 fperrad joined, dip left, cotto left
kthakore Happy Holidays!!! 16:10
fbrito :) 16:21
Thank you
yeah! aloha is back
kthakore fbrito: :D! 16:27
16:30 fbrito1 joined 16:36 fbrito left
dalek TT #1903 created by plobsing++: tailcall method invocation pops exception handlers 16:38
TT #1903: trac.parrot.org/parrot/ticket/1903
16:40 Matt_ left 16:55 theory joined 16:59 ambs left 17:26 Matt_ joined
dalek Some 'nwellnhof' person just gave me a 258585 byte commit packet. They need to be more careful in the future. 17:53
fbrito1 Wow :o 17:57
mikehh we definately need to change that message and how dalek handles branch merges 18:12
sorear I don't think we have much choice in the latter - github just gives us a bunch of commits to display 18:14
I'm welcome to suggestions though
as for the first, what so you suggest instead? 18:15
cxreg github gives you what format?
18:18 contingencyplan joined
mikehh we probably need to add a filter so it just shows the commit message for the merge 18:20
and the committer
sorear cxreg: help.github.com/post-receive-hooks/ 18:23
mikehh: finding "the commit message for the merge" will require some heuristics 18:24
cxreg oh, i didnt realize github had postbacks 18:25
mikehh Add minimum GC threshold option | nwellnhof++ | commit 9b2b65645c30756f4846 18:27
18:30 theory_ joined, theory left, theory_ is now known as theory
mikehh sorear: actually there seems to be two commits there 1st - Merge branch 'master' into nwellnhof/gc_dynamic_threshold -> Showing 547 changed files with 17,356 additions and 3,729 deletions. (two conflicts) 18:41
followed by the Add minimum GC threshold option -> Showing 18 changed files with 144 additions and 188 deletions 18:42
sorear mikehh: 1. how did you get a copy of the commit packet? 2. so the 258kb was mostly just listing changed files? 18:44
18:45 fbrito1 left
mikehh went to github parrot/parrot - selected branch nwellnhof/gc_dynamic_threshold - looked at commit history 18:49
18:50 cottoo is now known as cotto
mikehh of course it also has the commits in the history that have just been merger as well 18:51
merged 18:52
sorear looking at the raw packet it contains 1 commit (Add minimum GC threshold option) + 1 commit (Merge branch...) + many commits (The contents of the merge) 18:53
mikehh it looks like nwellnhof merged and then fixed the conflicts (and maybe other things) before pushing 18:54
dalek m-eta-wink-kzd: 21a60fc | plobsing++ | src/Ī©Ī·.coda.winxed:
properly return in :immediate sub
 19:02
m-eta-wink-kzd: 7a6bffc | plobsing++ | src/ometa-base.winxed:
provide rudimentary parse failure support
cotto ~ 19:06
19:27 Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 19:54 rfw joined 20:08 Matt_ left
dukeleto howdy 20:21
Yuki`N: pong
20:26 Matt_ joined
cotto hi dukeleto 20:29
dukeleto cotto: how goes? Has the gci bottleneck been taken care of?
cotto yup 20:30
dukeleto works on his TPF grant blog post 20:37
leto.net/dukeleto.pl/2010/12/parrot...ate-1.html 21:19
21:34 rurban_ joined 21:37 rurban left, rurban_ is now known as rurban 21:58 theory_ joined, theory left, theory_ is now known as theory 22:02 theory left 22:03 Matt_ left 22:04 theory joined 22:22 fperrad left 22:32 bacek left 22:34 nwellnhof joined 22:47 theory left 22:50 theory joined
moderator Parrot 2.11.0 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Onward and upward with Google Code-In | Please test all nwellnhof/* branches | Merge html_cleanup 23:03
nwellnhof hijacks the topic 23:04
23:06 theory left, theory joined 23:10 jan left 23:17 particle1 joined 23:18 Yuki`N joined 23:21 particle left 23:32 M_o_C joined
nwellnhof Rakudo build time on a 768MB virtual machine: 23:41
master: 4m56.880s
nwellnhof/gc_dynamic_threshold: 4m11.929s 23:42
sorear nwellnhof: any prayer of success with 384? 23:44
nwellnhof 384?
MB? 23:45
dukeleto nwellnhof: yes
nwellnhof: or 256mb, even
nwellnhof my branch should give even more performance improvements with 512 or 384 MB. 23:46
256 MB is probably not enough to build Rakudo without excessive swapping. 23:47
if run you OOM, my branch doesn't help. 23:48
well, you could lower the dynamic threshold. that might help to build Rakudo in some corner cases. 23:51
in nwellnhof/gc_dynamic_threshold there are now two command line options to tune the GC: 23:54
--gc-min-threshold=KB: the minimum amount of memory the GC will allocate before running a GC. Basically equivalent to the current static threshold. Default is 1/8 of system memory like in master. 23:55
--gc-dynamic-threshold: percentage of actually needed memory that will be allocated before a GC is run. this is only used if it is larger than min-threshold. 23:57
23:59 jan joined

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!